clipka <ano### [at] anonymousorg> wrote:

> >> No. No. Absolutely no. Unless you /specifically/ want to model something
> >> that glows in the dark, do NOT EVER use "emission".
> >>
> >> If you want to brighten up shadows, use radiosity or "ambient".


[Kenneth]:
> > That's interesting; I thought ambient and emission were identical in their
> > effect (in a non-radiosity scene, that is.) Basically that they are
> > interchangable.

[Clipka}:
> And that's /exactly/ the reason: From their effect in /non-radiosity/
> scenes they are interchangable, but from both their /intention/ and
> their effect in /radiosity/ scenes the two are worlds apart.
>
> Using "emission" to model indirect illumination of shadowy areas so
> /utterly/ defies the purpose of the feature that I /must/ rant about it,
> as a matter of principle. Kind of like if you'd use "assumed_gamma 500"
> or something like that.

You're certainly allowed to rant now and then ;-) Your general patience in
answering questions is one of the best qualities of the newsgroups.

Hmm, I haven't yet tried assumed_gamma 500; sounds interesting...  :-P

> Next in the triad came radiosity, which serves /exactly/ the same
> purpose, albeit with much higher quality. So theoretically, the
> "ambient" mechanism and radiosity shouldn't be used together.

Yep, that's clear.

> However, people figured out that radiosity still picked up the light
> from the "ambient" mechanism...
[snip]

Ah, I get it now. In a nutshell, and speaking simplistically, it's from the
desire to not have to change a lot of code in an already-coded scene, when
switching from non-radiosity to using it. For a scene that is not intended for
radiosity effects-- and I write a fair number of such scenes -- ambient IS the
logical way to go. But in case I change my mind later, I sometimes write
my object finishes{...} like this:
                    ambient .1
                    emission 0
                    diffuse *something* (but not 1.0, of course)
(or as a #default finish)
..... in case I want to run radiosity later--- which might then require me to
change a few emission values, *if* I decide to make any of those objects 'glow'
with radiosity light. Like a LDR sky sphere/image, for example. (Which
admittedly doesn't concern an object's 'ambient shadow' lighting, which is the
crux of the argument here.) But when rendering that sky sphere in my (original)
NON-radiosity version of the scene, I should choose ambient 1.0 for it at first
(according to the argument), assuming that I want to actually see its nicely-lit
image in the render. That will produce a non-fatal warning message of course--
"ambient is too high." But I ignore that, naturally.

Here's the rub, though: If, when originally writing the scene, I imagine that I
*might* eventually include radiosity, the finish for my sky sphere would be
easier/more logical to write as...
                    ambient 0
                    emission 1.0
                    diffuse 0
.... since ambient and emission have the same effect without radiosity. In other
words, that set-up would appear to be satisfactory for *both* worlds. Yet it
definitely goes against the (all-inclusive?) argument of not using emission in a
non-radiosity environment.

Post a reply to this message